Detekce objektů v mračnu 3D bodů

Abstract

Self-driving cars and robots have became a hot topic for a broad audience in a recent years. Robot has to know map of the environment and their location in such a map for proper path planning and obstacle detection. In unknown environment the robot doesn’t have the map so it can’t find itself in it. This situation is solved by Simultaneous Localization and Mapping, which is a long name for SLAM. In this thesis you will learn what SLAM actually means and how to solve it. Laser sensor Velodyne LiDAR produces point cloud and is used for creating the map. Point cloud is fed into open-source software Google Cartographer. In the thesis its configuration is described and obstacle detector is implemented for the created map.